Extraction of Proanthocyanidins from Chinese Wild Rice (Zizania latifolia) and Analyses of Structural Composition and Potential Bioactivities of Different Fractions

Due to the importance of proanthocyanidin bioactivity and its relationship with chemical structure, ultrasound-assisted extraction and purification schemes were proposed to evaluate the proanthocyanidin content and analyze the structural composition and potential bioactivities of different proanthocyanidin fractions from Chinese wild rice (Zizania latifolia). Following an optimized extraction procedure, the crude wild rice proanthocyanidins (WRPs) were purified using n-butanol extraction, chromatography on macroporous resins, and further fractionation on Sephadex LH-20 to yield six specific fractions (WRPs-1–WRPs-6) containing proanthocyanidin levels exceeding 524.19 ± 3.56 mg/g extract. Structurally, (+)-catechin, (−)-epicatechin, and (−)-epigallocatechin were present as both terminal and extension units, and (−)-epicatechin was the major extension unit, in each fraction. This is the first preparation of WRP fractions with a different mean degree of polymerization (mDP), ranging from 2.66 ± 0.04 to 10.30 ± 0.46. A comparison of the bioactivities of these fractions revealed that fractions WRPs-1−WRPs-5 had significant DPPH radical scavenging activities, whereas fraction WRPs-6 with a high mDP showed better α-glucosidase and pancreatic lipase inhibitory effects. These findings should help define possible applications of WRPs to functional foods or nutraceuticals.


Introduction
Wild rice, the seed of the aquatic plant Zizania (family Poaceae), has long been recognized as a health-promoting whole grain, particularly because of its health benefits, including the ability to suppress oxidative stress, reduce hyperlipidemia, and prevent atherogenesis, type 2 diabetes, and obesity [1][2][3]. Since wild rice is gluten-free and safe for human consumption, the growing interest in wild rice has increased its commercialization in North America [4]. There are four known species of wild rice. Of these, Z. aquatica, Z. palustris, and Z. texana are indigenous to North America, whereas Z. latifolia is native to China, Japan, and Vietnam [5]. Chinese wild rice (Z. latifolia), which is widely distributed in areas along the Yangtze and Huai Rivers, is an age-old food that has been traditionally used to treat a variety of ailments in Chinese medicinal practice [2,6]. Wild rice is a rich source of phenolic Since extraction is the most important step to recover the highest amount of the target compounds from plants [17], a preliminary single factor experiment was conducted to determine the effects of operation parameters of ultrasound-assisted extraction (UAE) on the content of WRPs. According to the experimental results ( Figure S1 in the Supplementary Materials), the four most important independent variables at three levels, namely the concentration of aqueous ethanol (80, 90, and 100%; v/v), the liquid-solid ratio (40, 50, and 60 mL/g; v/w), the extraction temperature (30,40, and 50 • C), and the ultrasonic power (300, 350, and 400 W), were selected for optimizing the extraction conditions using RSM based on a Box-Behnken design (BBD).

Fitting the Model
The corresponding experimental results for the content of WRPs of each run in BBD are presented in Table S1 in the Supplementary Materials. By applying multiple regression analysis on the experimental data, the equation describing the correlation between WRPs content and the four variables was as follows: Molecules 2019, 24, 1681 3 of 13 Y = 6.09 + 0.80X 1 − 0.10X 2 − 0.098X 3 − 0.075X 4 − 0.014X 1 X 2 − 0.30X 1 X 3 − 0.14X 1 X 4 − 0.043X 2 X 3 − 0.26X 2 X 4 + 0.19X 3 X 4 − 1.37X 1 2 where Y is the WRPs content (mg/g rice); X 1 , X 2 , X 3 , and X 4 represent the coded variables for the concentration of aqueous ethanol, the liquid-solid ratio, the extraction temperature, and the ultrasonic power, respectively. The data for the analysis of variance (ANOVA) statistical test of the model are shown in Table 1. The small p-value (p < 0.0001) suggested that the model was highly significant. Meanwhile, a lack of fit F-value of 2.88 and an associated p-value of 0.1596 implied that there was an excellent agreement between the experimental values and the predicted values. In addition, a determination coefficient (R 2 ) of 0.9941 and an adjusted determination coefficient (adj R 2 ) of 0.9882 indicated that there was a satisfactory correlation between the experimental WRPs content and the values predicted by the equation. A high degree of precision and good reliability of the data were demonstrated by a low coefficient of variation (CV = 2.24%). These results suggested that the model works well at predicting WRPs extraction.

Optimization of Extraction Conditions
Three-dimensional surface plots were used to explain the interactions of the variables and to determine the optimal level of each independent variable to produce a maximal response. When the extraction temperature and ultrasonic power were set at zero level, the WRPs content increased initially as the concentration of aqueous ethanol and the liquid-solid ratio increased and then slightly decreased ( Figure 1A). Similar effects on WRPs content of other variables are shown in Figure 1B-F. The results were in accordance with the single factor experiment and the ANOVA analysis. The maximum predicted value of WRPs content (6.28 mg/g rice) could be obtained under the following optimal conditions: a concentration of aqueous ethanol of 93.73%, a solid-liquid ratio of 50.07 mL/g, an extraction temperature of 44.39 • C, and an ultrasonic power of 330.34 W. Considering the operability in actual processing procedures, a verification experiment was carried out under the following modified conditions: a concentration of aqueous ethanol of 94%, a solid-liquid ratio of 50 mL/g, an extraction temperature of 44 • C, and an ultrasonic power of 330 W. Under these conditions, the WRPs content of the extract from the verification experiment was 6.33 ± 0.10 mg/g rice, which was in good agreement with the predicted value.

Proanthocyanidin Content and in Vitro Bioactivities of the Isolated Fractions
The crude WRPs obtained under the optimal UAE condition were first divided into four fractions soluble in n-hexane, ethyl acetate, n-butanol, and water, respectively. The proanthocyanidin content, DPPH radical scavenging activity, and α-glucosidase inhibitory activity were evaluated to determine the active fractions. As shown in Figure 2A-C, the n-butanol fraction had the highest proanthocyanidin content, DPPH radical scavenging activity, and α-glucosidase inhibitory activity.
The active n-butanol fraction was then further loaded onto a D101 macroporous adsorption resin column to obtain four fractions (fractions C1-C4), which were screened for their proanthocyanidin content, antioxidant activity, as well as α-glucosidase and pancreatic lipase inhibitory activities. As shown in Figure 2D-F, the fractions C2 and C3, with the highest proanthocyanidin contents of 282.24 ± 2.07 and 112.80 ± 1.55 mg/g extract, respectively, displayed good DPPH radical scavenging activity and α-glucosidase inhibitory activities. The proanthocyanidins were, therefore, mainly distributed in fractions C2 (10-30% ethanol effluents) and C3 (40-60% ethanol effluents), which were combined and concentrated for next fractionation. All the four fractions at a concentration of 10 mg/mL showed no obvious inhibitory activity against pancreatic lipase.

Proanthocyanidin Content and In Vitro Bioactivities of the Isolated Fractions
The crude WRPs obtained under the optimal UAE condition were first divided into four fractions soluble in n-hexane, ethyl acetate, n-butanol, and water, respectively. The proanthocyanidin content, DPPH radical scavenging activity, and α-glucosidase inhibitory activity were evaluated to determine the active fractions. As shown in Figure 2A-C, the n-butanol fraction had the highest proanthocyanidin content, DPPH radical scavenging activity, and α-glucosidase inhibitory activity.
The active n-butanol fraction was then further loaded onto a D101 macroporous adsorption resin column to obtain four fractions (fractions C1-C4), which were screened for their proanthocyanidin content, antioxidant activity, as well as α-glucosidase and pancreatic lipase inhibitory activities. As shown in Figure 2D-F, the fractions C2 and C3, with the highest proanthocyanidin contents of 282.24 ± 2.07 and 112.80 ± 1.55 mg/g extract, respectively, displayed good DPPH radical scavenging activity and α-glucosidase inhibitory activities. The proanthocyanidins were, therefore, mainly distributed in fractions C2 (10-30% ethanol effluents) and C3 (40-60% ethanol effluents), which were combined and concentrated for next fractionation. All the four fractions at a concentration of 10 mg/mL showed no obvious inhibitory activity against pancreatic lipase.

Structural Composition of Fractions WRPs-1-WRPs-6
The reversed-phase UPLC chromatograms of fractions WRPs-1-WRPs-6 ( Figure S2 in the supplementary materials) implied that they have distinctly different chemical compositions. The structural composition of fractions WRPs-1-WRPs-6 was determined by using phloroglucinolysis,  Data are expressed as mean ± SD (n = 3). Values with different letters in the same column are significantly different (p < 0.05).

Structural Composition of Fractions WRPs-1-WRPs-6
The reversed-phase UPLC chromatograms of fractions WRPs-1-WRPs-6 ( Figure S2 in the Supplementary Materials) implied that they have distinctly different chemical compositions. The structural composition of fractions WRPs-1-WRPs-6 was determined by using phloroglucinolysis, which has been proved to be an efficient method for structure analysis of proanthocyanidins [18,19]. In the degradation reaction, terminal units of proanthocyanidins are liberated as free flavan-3-ol monomers, while the extension units, released as electrophilic intermediates, are attacked by the nucleophile (phloroglucinol) to form the corresponding phloroglucinol adducts. The composition of the flavan-3-ol units and the mDP of the proanthocyanidins were determined by UPLC-LTQ-Orbitrap-MS analysis of the cleavage products. Figure 3 shows the reversed-phase UPLC chromatograms of the fractions WRPs-1-WRPs-6 after phloroglucinolysis. The major cleavage products detected were identified by comparing their retention times and MS data with those of authentic standards or literature data [11,20], and results are listed in Table 3. The WRPs were mainly composed of (+)-catechin (C), (−)-epicatechin (EC), and (−)-epigallocatechin (EGC) as terminal and extension units. The proportions of constitutive units and mDP of proanthocyanidins in each fraction were also calculated ( Table 4). The terminal units of fractions WRPs-1-WRPs-6 were mainly C and EC with proportions of 4.21 ± 0.12-15.15 ± 0.30% and 3.80 ± 0.06-11.83 ± 0.42%, respectively. With respect to the extension units, EC with increased contents, from fractions WRPs-1 to WRPs-6, had the highest proportions (44.03 ± 0.53-70.63 ± 0.69%), followed by C (15.79 ± 0.21-22.89 ± 0.26%), suggesting EC to be the major extension unit of WRPs. Compared with C and EC, EGC was found at much lower levels in the terminal and extension units (1.70 ± 0.04-9.55 ± 0.33% and 2.87 ± 0.08-3.87 ± 0.10%, respectively). The mDP gradually increased from fractions WRPs-1 to WRPs-6, with fraction WRPs-1 having a minimum of 2.66 ± 0.04 and fraction WRPs-6 having a maximum of 10.30 ± 0.46. nucleophile (phloroglucinol) to form the corresponding phloroglucinol adducts. The composition of the flavan-3-ol units and the mDP of the proanthocyanidins were determined by UPLC-LTQ-Orbitrap-MS analysis of the cleavage products. Figure 3 shows the reversed-phase UPLC chromatograms of the fractions WRPs-1-WRPs-6 after phloroglucinolysis. The major cleavage products detected were identified by comparing their retention times and MS data with those of authentic standards or literature data [11,20], and results are listed in Table 3. The WRPs were mainly composed of (+)-catechin (C), (−)-epicatechin (EC), and (−)-epigallocatechin (EGC) as terminal and extension units. The proportions of constitutive units and mDP of proanthocyanidins in each fraction were also calculated ( Table 4). The terminal units of fractions WRPs-1-WRPs-6 were mainly C and EC with proportions of 4.21 ± 0.12-15.15 ± 0.30% and 3.80 ± 0.06-11.83 ± 0.42%, respectively. With respect to the extension units, EC with increased contents, from fractions WRPs-1 to WRPs-6, had the highest proportions (44.03 ± 0.53-70.63 ± 0.69%), followed by C (15.79 ± 0.21-22.89 ± 0.26%), suggesting EC to be the major extension unit of WRPs. Compared with C and EC, EGC was found at much lower levels in the terminal and extension units (1.70 ± 0.04-9.55 ± 0.33% and 2.87 ± 0.08-3.87 ± 0.10%, respectively). The mDP gradually increased from fractions WRPs-1 to WRPs-6, with fraction WRPs-1 having a minimum of 2.66 ± 0.04 and fraction WRPs-6 having a maximum of 10.30 ± 0.46.      Upon comparing the aforementioned information with the bioactivities of fractions WRPs-1-WRPs-6, it is clear that the structure of WRPs greatly affects their bioactivity. The DPPH radical scavenging activity was the lowest in fraction WRPs-6 with the largest mDP, followed by fraction WRPs-1 with the smallest mDP, indicating that there was an increase and then a fall in DPPH radical scavenging activity as the mDP increased. Similarly, Jerez et al. found that for proanthocyanidins from the bark of Pinus radiata, there was an increase in DPPH radical scavenging activity up to 6.5 mDP and then a fall in DPPH scavenging activity as mDP further increased (9.2-14.6 mDP) [21]. With respect to enzyme inhibitory activities, fraction WRPs-6, which had the largest mDP, showed the highest inhibitory activities against α-glucosidase and pancreatic lipase. Previous reports also suggested that the α-glucosidase inhibitory activity of proanthocyanidins from the leaves of Chamaecyparis obtusa var. formosana and pancreatic lipase inhibitory activity of proanthocyanidins from the fruits of Diospyros kaki both increased with mDP [22]. Therefore the mDP of WRPs is a significant determinant of their promising potential to be used as natural food antioxidants and enzyme inhibitors.

Plant Materials and Chemicals
Wild rice Z. latifolia was collected from Jiangling County, Jingzhou City, Hubei Province, China (30 • 13 10 N; 112 • 34 5 E), in September 2017. The sample was obtained by manually harvesting mature plant tassels, drying and then dehulling them to obtain the seeds. The seeds were ground to a fine powder in a mechanical grinder and then sieved through a 0.45 mm sifter.

Ultrasound-Assisted Extraction Procedure
The WRPs were extracted using a KQ-2200DB ultrasonic cleaning bath (Kunshan Co., Ltd., Kunshan, China) and the temperature was preset before the extraction process. After extraction, the mixture was centrifuged at 3000× g for 20 min, and the combined supernatants were used as the crude WRPs to determine the proanthocyanidin content.

Experimental Design
According to previous studies [23,24], a single factor experiment was firstly performed to investigate the effects of the concentration of aqueous ethanol (60, 70, 80, 90, and 100%; EtOH %, v/v), the liquid-solid ratio (20,30,40,50, and 60 mL/g; v/w), the extraction temperature (30, 40, 50, 60, and 70 • C), the ultrasonic power (200, 250, 300, 350, and 400 W), and the extraction time (20,30,40,50, and 60 min) on the content of WRPs obtained. Subsequently, a BBD with four important variables at three levels, which were determined based on the single factor experimental results, was employed to optimize the extraction conditions. The combined effects of the four independent variables: the concentration of aqueous ethanol (EtOH %, X 1 ), the liquid-solid ratio (mL/g, X 2 ), the extraction temperature ( • C, X 3 ), and the ultrasonic power (W, X 4 ) were evaluated. The coded and uncoded (actual) levels of the independent variables are listed in Table S1 in the Supplementary Materials. The response variables were fitted to the following quadratic polynomial model: where Y represents the response variable, WRPs content, X i and X j are the independent variables affecting the response, and β 0 , β i , β ii , and β ij are the regression coefficients of the model (intercept, linear, quadratic, and interaction terms, respectively). The Design Expert Software (Version 8.0.6; Stat-Ease Inc., Minneapolis, MN, USA) was used for the experimental plan, data analysis, model generation, and determination of optimum conditions. The relationship between independent variables and responses was analyzed by response surface plots. Optimum conditions for UAE were calculated according to the desirability function.

Acid-Catalysis in the Presence of Phloroglucinol (Phloroglucinolysis)
Phloroglucinolysis of the isolated fractions WRPs-1-WRPs-6 was performed based on an earlier study [18]. Briefly, 5 mg of sample was reacted in 1 mL of 0.1 N HCl in methanol containing 50 g/L phloroglucinol and 10 g/L ascorbic acid in a stoppered test tube. The reaction mixture was heated at 50 • C for 20 min, after which 5 mL of 40 mM aqueous sodium acetate was added to stop the reaction. The cleavage products were immediately analyzed by UPLC-LTQ-Orbitrap-MS.
For mass detection, an electrospray ionization source was operated in positive mode with a scan range from m/z 150 to 1500. The capillary temperature was 350 • C. Nitrogen was used as the sheath gas and auxiliary gas, and the gas flow was set at 30 and 5 arbitrary units, respectively. The spray voltage was 4000 V. The collision energy was 35% to adjust for collision-induced dissociation for the best performance. The Xcalibur 2.1 software (Thermo Scientific, San Jose, CA, USA) was used for data analysis.
The identification of C, EC, EGC, and the corresponding phloroglucinol adducts was achieved by comparing their UPLC retention times and MS data with those of authentic standards or literature data. C, EC, and EGC were quantified based on their standards. For quantitation of the phloroglucinol adducts, the respective flavan-3-ol monomers were used as the standards. With this procedure, the phloroglucinol adducts were assumed to have the same molar absorptivities as their respective flavan-3-ol monomers [18,28]. To calculate the mDP of the proanthocyanidins, the sum of all subunits (flavan-3-ol monomers and phloroglucinol adducts, in moles), was divided by the sum of all flavan-3-ol monomers (in moles).

Determination of Proanthocyanidin Content
The proanthocyanidin content was measured using a modified vanillin-H 2 SO 4 method [29]. Briefly, 20 µL of the sample was mixed with 100 µL of vanillin methanol solution (30 g/L) and 100 µL of H 2 SO 4 methanol solution (30%, v/v). (+)-Catechin was used as a reference. After incubation at room temperature for 5 min in the dark, the absorbance at 500 nm was measured using a Multiskan GO microplate spectrophotometer (Thermo Scientific, San Jose, CA, USA). The results for the crude WRPs and the isolated fractions were expressed as milligram (+)-catechin equivalents per gram of rice (mg/g rice) and milligram (+)-catechin equivalents per gram of extract (mg/g extract), respectively.

DPPH Radical Scavenging Assay
The DPPH radical scavenging activity was evaluated using the method described by Yuan et al. [30], with ascorbic acid as the positive control. The percentage of scavenging was calculated according to the following equation: where A sample represents the absorbance of the DPPH solution in the presence of the sample and A control is the absorbance of the DPPH solution without sample (which was replaced by the solvent). The corresponding IC 50 value, defined as the concentration of extract required to scavenge 50% of the DPPH radicals, was also calculated, and the result was expressed as micrograms extract per milliliter solvent (µg/mL).

α-Glucosidase Inhibition Assay
The α-glucosidase inhibition assay was carried out according to a previously reported method [31], with some modifications. Briefly, 10 µL of the test sample in DMSO was mixed with 620 µL of phosphate buffered saline (PBS; 0.1 M, pH 6.9) after which 10 µL of α-glucosidase (2 U/mL) in PBS was added and incubated at 37 • C for 10 min. Following this, 200 µL of 6 mM p-nitrophenyl α-D-glucopyranoside in PBS was added and incubated at 37 • C for another 20 min. The reaction was terminated by adding 1 mL of 1 M sodium carbonate aqueous solution and the absorbance at 400 nm was measured. Acarbose was used as the positive control. The percentage inhibition was calculated as follows: where A sample represents the absorbance of the test sample, A control is the absorbance of the control solution without sample (which was replaced by DMSO), and A background is the absorbance of the background solution without α-glucosidase (which was replaced by PBS). The IC 50 value (µg/mL) as similarly defined for the DPPH radical was also determined.

Pancreatic Lipase Inhibition Assay
The pancreatic lipase activity was measured using a previous method [31], with some modifications. Briefly, 10 µL of the test sample in DMSO was added to 500 µL of Tris-HCl buffer (0.1 M, pH 8.0), after which 200 µL of pancreatic lipase (2 mg/mL) in Tris-HCl buffer was added and incubated at 37 • C for 15 min. Following this, 200 µL of 12.5 mM p-nitrophenyl butyrate in Tris-HCl buffer was added and incubated at 37 • C for 20 min. The absorbance at 400 nm was measured and orlistat was used as the positive control. The percentage inhibition of pancreatic lipase was calculated as follows: where A sample represents the absorbance of the test sample, A control is the absorbance of the control solution without the sample (which was replaced by DMSO), and A background is the absorbance of the background solution without pancreatic lipase (which was replaced by Tris-HCl). The IC 50 value (µg/mL) was calculated.

Statistical Analysis
All results are the averages of at least three assay replicates and are expressed as the mean ± standard deviation (SD). Statistical analysis was performed using the SPSS 19.0 software (SPSS Inc., Chicago, IL, USA). p-Values < 0.05 were considered to be statistically significant.

Conclusions
In the present study, an optimized UAE was applied to extract proanthocyanidins from Chinese wild rice with high yield. Six specific fractions with mDP in increasing order were obtained after partition, purification, and fractionation of crude WRPs. Fractions WRPs-1-WRPs-5 possessed excellent DPPH radical scavenging activity, and fraction WRPs-6 with the highest mDP had α-glucosidase and pancreatic lipase inhibitory effects. These findings provide, for the first time, useful information on the structures and potential bioactivities of different WRP fractions. The high proanthocyanidin contents and the potential role of the different WRP fractions in oxidation resistance and management of diabetes and obesity make them worthy of further consideration as functional food ingredients or nutraceuticals. Future in vivo study on these beneficial effects will be carried out to facilitate the applications of WRPs.
Supplementary Materials: The following are available online, Table S1: Box-Behnken design with coded and actual values for the extraction conditions and response values for the content of wild rice proanthocyanidins. Figure S1: Effects of different extraction parameters on the content of wild rice proanthocyanidins in the single factor experiment. Figure

Conflicts of Interest:
The authors declare no conflict of interest.